PFENet icon indicating copy to clipboard operation
PFENet copied to clipboard

PFENet: Prior Guided Feature Enrichment Network for Few-shot Segmentation (TPAMI).

Results 38 PFENet issues
Sort by recently updated
recently updated
newest added

Hi~ thanks for your works! Previous works indicates 'the number of episodes' in code. However, when i run this code, the number of episodes looks like len(self.data_list) which depends on...

please teach me how to see a demo on windows Thank you

你好,我刚接触小样本分割,请教一下您,训练集和验证集是不是都包含同样的类别呀。

How to replace with my own dataset?

如题,链接已经失效了还是被删了

Hi, the links for pre-trained models is not working. Can you please share a new link? Thank you.

你好,首先感谢您分享的代码。train的时候,出现这样的错误 ![屏幕截图 2024-07-11 213744](https://github.com/dvlab-research/PFENet/assets/175322144/f8efddab-8abc-4264-8252-11487669edec) 经过debug,发现是 val_loader里的val_data里没有数据,这到底是怎么回事呢?用的是split=0

作者您好,向您请教下关于数据集划分的细节 我们知道验证集用于调整模型参数,测试集用于评估模型最终性能。 以PASCAL为例,交叉验证以类别为单位划分5个类别作为测试集,剩下15个类别作为训练集。在训练阶段,同步在测试集中采样1000个样本作为验证,但在最终测试时,是否包含验证所用样本?具体是如何区分验证集与测试集的? 感谢答复!